Maximum covariance analysis to identify intraseasonal oscillations over tropical Brazil

A reliable prognosis of extreme precipitation events in the tropics is arguably challenging to obtain due to the interaction of meteorological systems at various time scales. A pivotal component of the global climate variability is the so-called intraseasonal oscillations, phenomena that occur between 20 and 100 days. The Madden–Julian Oscillation (MJO), which is directly related to the modulation of convective precipitation in the equatorial belt, is considered the primary oscillation in the tropical region. The aim of this study is to diagnose the connection between the MJO signal and the regional intraseasonal rainfall variability over tropical Brazil. This is achieved through the development of an index called Multivariate Intraseasonal Index for Tropical Brazil (MITB). This index is based on Maximum Covariance Analysis (MCA) applied to the filtered daily anomalies of rainfall data over tropical Brazil against a group of covariates consisting of: outgoing longwave radiation and the zonal component u of the wind at 850 and 200 hPa. The first two MCA modes, which were used to create the $${ MITB}_1$$ and $${ MITB}_2$$ indices, represent 65 and 16 % of the explained variance, respectively. The combined multivariate index was able to satisfactorily represent the pattern of intraseasonal variability over tropical Brazil, showing that there are periods of activation and inhibition of precipitation connected with the pattern of MJO propagation. The MITB index could potentially be used as a diagnostic tool for intraseasonal forecasting.